projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 2a2bde4) | patch
R600/SI: Don't store scratch buffer frame index in MUBUF offset field
authorTom Stellard <thomas.stellard@amd.com>
Tue, 20 Jan 2015 17:49:43 +0000 (17:49 +0000)
committerTom Stellard <thomas.stellard@amd.com>
Tue, 20 Jan 2015 17:49:43 +0000 (17:49 +0000)
commita978a481bbe66becb01f4955b8e002e6cd59be23
tree2f1c4a8433eddbb78e8e27ac39430abeac10c85dtree | snapshot
parent2a2bde40ab702b78f4f78777a04e3b60bb4dbb2ccommit | diff
R600/SI: Don't store scratch buffer frame index in MUBUF offset field

We don't have a good way of legalizing this if the frame index offset
is more than the 12-bits, which is size of MUBUF's offset field, so
now we store the frame index in the vaddr field.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@226584 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/R600/AMDGPUISelDAGToDAG.cpp diff | blob | history
test/CodeGen/R600/scratch-buffer.ll [new file with mode: 0644] blob
C/C++ LLVM-based compilers that forbids OOTA behaviors